CT系列图像的三维层次化子块生长分割.pdfVIP

182 2011 ,47(11) Computer Engineering and Applications计算机工程与般用 CT系列图像的主维层次化子块生伏分割 黄展鹏 1，蒋世忠 1，鲍苏苏 2，黄欣甜 1 HUANG Zhanpeng l , JIANG Shizhong 1 , BAO Susu2, HUANG Xinzhe J 1.广东药学院医药信息工程学院，广州 510006 2.华南师范大学计算机学院，广州 510631 1.College of Medical Information Engineering. Guangdong Pharmaceutical University. Guan萨hou 510006 ,China 2.College of Compu阳， Sou伽 China Normal Universi纱，Guan萨hou 510631. china HUANG Zhanpeng , JlANG Shizhong, BAO SU8U , et aLCT image I吨mentation based on three-dfmensional hierarchical lub-block region growing.Computer Engineering and Applications , 1011 , 47( 11) : 182-184. Abs仿制: Se阴阳tation of region of 恤阳est based on abdonúnal Compu时 Tomogr叩，hy ( CT) sequences images is a crucial 阳p in surgical planning.However.严ecisely carrying out 也is step remains a challenge.By considering the siωation of ke叩 切g the wrong boundary as 出e liver surface. an approach for CT image se伊lentation based on 3D hierarchical region grow- ing algori也m is pr叩osed.1n the 叩proach ， severa1 seed points are selected flrstly.Considering the similarity of the neighbor slices in 也e medical volume data,a modifled 3D sub--block region growing algorithm is developed.1n orderωreduce theω- erin阳刚on ， the growing criteria are estimated auωmatically 由rough investigation of 幽刷刷tical characteristics in the 10翩 cal regions of the seed points.Experiments results show the proposed method can efficiently se伊lent the live region 食om se幽 rial abdominal CT images with little user interaction. Key words: 3D region grow切g algori伽m;hierarchical segmenta悦。n;live川CT 伽mge-s叫uences 摘 要:准确地从CT系列图像提取感兴趣的组织是手术规划的基础，针对肝脏轮廓分割存在分割不会的问题，提出丁暴于三维 区域生长算法的腹部CT阁像分割方法.算法首先由用户选择若干个生长点，然后克分利用 CT系列图像层间的相似性，提出基 于子块的改进区域生长算法，实现三维的层次化子块民域生长，以是准确提取肝脏区域，其中生长准则由系统分析用户选择的生 长点的部城子块属性获得，以减少用户的干预.实验结果表明，算法能在较少的干预下快速分割出来CT系列困像中的肝剧告廓. 关键询:三维区域生长算法;层次化分割;肝脏;CT 系列图像 ∞1: 10.3778/j.?ssn.1002-833 1.201 1.11.051 义意楠号: 1002-8331(2011) 11-0182-的 文献标识码:A 中回分费号:TP391. 1 引宵 癌在是今天最常见的发病和死亡原因之一，据世界卫生 组织统计，才是世界每年布 l ∞0多万新病例和6∞多万人死L:. 其中